A recent study of selective attention operations in schizophrenia reported subtle right visual field (RVF) hemineglect in a sample of non- chronic schizophrenics (Posner et al, 1988). This finding stimulated speculation about specific lateralized lesions (left parietal and/or left anterior) and attentional operations (i.e., diminished left hemisphere ability to generate preparatory sets) in schizophrenia. These findings were also used to explain other attention abnormalities in schizophrenia (i.e., vigilance). Replication of this important finding is needed to ensure that model building activity in selective attention is on solid ground. Also, determination of the mechanism underlying the finding is needed for accurate neuroanatomical theory and further research on its diagnostic or etiologic value. Specifically, the hemineglect may be caused by a failure in stimulus selection, i.e., filtering, or an inability to generate attentional gradients, which play a key role in generating preparatory, expectational sets. Finally, the temporal stability of deficits in selective attention have not been established and their relation to neuroleptic medication and clinical state have yet to be defined. This project will replicate the Posner et al. (1988) study and assess performance on a shape identification, i.e., filtering, task (LaBerge and Brown, 1989) through a four group, repeated measures design (N=120; baseline, 3 and 6 month assessments). Chronic schizophrenics on one of three neuroleptic regimens (clozapine, risperidone, prolixin decanoate, n's=30) and normal controls (n=30) will be followed for six months. Patients will remain on the same neuroleptic throughout the study. Serial assessments will also be taken of psychiatric symptoms, cognitive status, and gross hemineglect. The three schizophrenic groups will be matched on duration of illness. All four groups will be matched on age and gender. The hemineglect hypothesis will be tested along with tests of an early filtering mechanism for visual shape, and the ability to generate attentional gradients.